Protein expression and purification
The cDNA sequence encoding an engineered PPR domain (residues 84-292 with solubilizing substitutions Y266N, F284Q, and F291Q) was subcloned into the pSMT3 vector (kindly provided by Christopher Lima, Memorial Sloan Kettering Cancer Center), which encodes an N-terminal His6-SUMO tag. The engineered PPR domain was expressed in E. coli strain BL21- CodonPlus (DE3)-RIL (Agilent Technologies) at 18 °C overnight after induction with 0.5 mM IPTG. The cells were collected by centrifugation, and pellets were resuspended in lysis buffer (containing 50 mM Tris-HCl, pH 8.0, 500 mM NaCl) and stored at −80 °C until use. The cells were disrupted by sonication followed by centrifugation to remove cell debris. The soluble fraction was applied to a Ni-NTA agarose column and thoroughly washed with lysis buffer containing 20 mM imidazole. The target SUMO fusion protein was eluted with lysis buffer containing 400 mM imidazole. The fusion protein was cleaved overnight with 0.2 mg of Ulp1 protease and dialyzed against a buffer containing 50 mM Tris-HCl, pH 8.0, 150 mM NaCl and 0.5 mM TCEP. The protein was then loaded onto a HiTrap Q column (GE Healthcare), which did not bind PRORP1 (Figure 3-1A). The flow-through fractions were pooled and dialyzed against a buffer containing 25 mM Hepes-NaOH, pH 7.0, 150 mM NaCl and 0.5 mM TCEP. The sample was loaded onto a HiTrap SP column (GE Healthcare). Bound proteins were eluted using a linear gradient from 0.1 to 1 M NaCl in 25 mM Hepes-NaOH, pH 7.0. Peak fractions containing the engineered PPR domain were pooled and concentrated and reducing agent was added (final concentration of 1 mM dithiothreitol or 5 mM 2-mercaptoethanol). The protein was purified further using a Superdex 75
10/300 GL column (GE Healthcare), equilibrated with 25 mM HEPES, pH 7.0, 200 mM NaCl, and 0.5 mM TCEP. Final purified protein was concentrated to 5 mg/mL.
Figure 3-1 Purification of the PRORP1 PPR domain and PRORP1 PPR domain-tRNAPhe complex for crystallization.
(A) Chromatography profiles and SDS-PAGE gels for purification of the PRORP1 PPR domain. For all chromatograms, fractions in the yellow highlighted region were pooled and loaded onto the subsequent column. For all gels, L marks the lane with the protein sample loaded onto the column and FT marks the lane with a representative flow through protein sample. The cleaved SUMO- PRORP1 PPR domain fusion protein preparation was loaded onto a HiTrap Q column to remove E. coli nucleic acid contamination, which bound to the column and was eluted in complex with the PRORP1 PPR domain (purple region) or as free nucleic acid (pink region). The majority of SUMO protein and PRORP1 PPR domain flowed through (yellow region). The HiTrap Q flow through fractions were pooled and loaded onto a HiTrap SP column and eluted with a 10-100% gradient of 1M NaCl. The PRORP1 PPR domain eluted within a single peak (yellow region). Much of the SUMO protein did not bind to the HiTrap SP column (see FT fraction lane). The pooled HiTrap
SP peak fractions were loaded onto a Superdex 75 column, which separated the PRORP1 PPR domain (yellow region) from residual SUMO protein (green region). (B) Chromatography profile, SDS-PAGE gel, and TBE-PAGE gel for purification of the PRORP1 PPR domain-tRNAPhe complex. The fractions in the red region were pooled and concentrated to A260 =43 for crystallization.
For crystallization, the engineered PPR domain was mixed with commercially available yeast tRNAPhe (Sigma-Aldrich). Protein and RNA were mixed at a ratio of 1:1.05, and the mixture was incubated at 4 °C overnight. The protein-RNA complex was purified further using a Superdex 75 10/300 GL column (GE Healthcare), equilibrated with 25 mM HEPES, pH 7.5, 150 mM NaCl, 2 mM MgCl2, and 0.5 mM TCEP (Figure 3-1B). Peak fractions containing the complex were pooled and concentrated to A260 = 43.
Full-length PRORP1 protein for binding assays was purified as described in a previous study (22). The protein was purified by Ni-NTA agarose chromatography, followed by purification on a HiTrap SP column. Peak fractions containing the full-length protein were pooled and concentrated. The protein was purified further using a HiLoad 16/60 Superdex 75 column (GE Healthcare) equilibrated with 20 mM MOPS, pH 7.8, 300 mM NaCl and 0.5 mM TCEP. The peak fractions were pooled and concentrated to 76 µM. Mutant proteins were expressed at equivalent levels to WT protein, behaved similarly during the purification steps, and no differences were detected in CD spectra.
Circular dichroism analyses
To assess folding of the mutant proteins, we measured circular dichroism (CD) spectra of full-length PRORP1 wild-type, R212K, R210S, and Y133Q proteins (Figure 3-2). The CD spectra were measured on a Chirascan CD spectrometer (Applied Photophysics) at 20 °C. For each sample (300 μl in a 0.1 cm light-path cell), three scans were accumulated in the wavelength range of 190– 260 nm with a 0.2 nm step size. Protein samples were 0.13 mg/ml in 25 mM sodium phosphate, pH 7.5, 200 mM NaCl. The raw CD data were adjusted by subtracting a buffer blank. CD spectra of wild-type and mutant proteins displayed negative ellipticities at 208/222 nm and 215 nm, which indicate the presence of α helices and β strands, respectively.
Crystallization, data collection, structure determination and refinement
Crystallization of the purified PRORP1 PPR domain-tRNAPhe complex was performed by the sitting-drop vapor diffusion method at 4 °C. Sitting drops contained 250 nL of protein-RNA complex solution mixed with 250 nL of reservoir solution (1.4-1.5 M sodium citrate). Prior to data collection, crystals were transferred to a cryoprotectant solution containing 1.6 M sodium citrate and flash cooled to -180 °C. X-ray diffraction data were collected at beamline 22-ID of the Advanced Photon Source (APS) at 100 K with a wavelength of 1.000 Å. The data were processed using HKL2000 (HKL Research Inc.) (41). Phases were determined by molecular replacement using the program Phaser and search models of the PPR domain of Arabidopsis PRORP1 (PDB ID: 4G24) and yeast tRNAPhe (PDB ID: 1EHZ). Model building was carried out with the program Coot (42). The programs Refmac5 (43) and Phenix.refine (44) were used for refinement. The structures displayed good geometry when analyzed by MolProbity (45). Approximately 98% and 2% of the residues constituting the PPR domain were in the most favored and allowed regions of the Ramachandran plot, respectively. Modified bases were modeled into the structure: 2-methyl-
guanosine at position 10, 5,6-dihydrouridine (D) at positions 16 and 17, N2-dimethylguanosine at position 26, O2´-methyl-cytidine at position 32, O2´-methyl-guanosine at position 34, the Y base or wybutosine at position 37, pseudouridine (ψ) at positions 39 and 55, 5-methyl-cytidine at positions 40 and 49, 7-methyl-guanosine at position 46, 5-methyl-uridine at position 54, and 1- methyl-adenosine at position 58. The average B factor for the tRNA is high due to poor electron density in regions of the tRNA that do not contact the PRORP1 protein. However, the electron density is strong at the tRNA D and TψC loops where it contacts PRORP1 (Figure 3-3A).
In vitro transcription
Pre-tRNAs were synthesized as previously described (36,46) through run-off transcription from a restriction-digested (BstNI) pUC18 plasmid encoding Bacillus subtilis pre-tRNAAsp. In vitro transcription reactions were run in 5:1 excess of 5´-O-monophosphorothiate guanosine (GMPS) to GTP. The 5´-GMPS pre-tRNA product was reacted with 5-iodoaceamidofluorescein (5-IAF) in a 1:40 molar ratio (RNA:5-IAF) to produce a 5´-fluorescein labeled product. Labeling reactions were carried out in 10 mM Tris, pH 7.2 with 1 mM EDTA for 16 h at 37 °C yielding 25- 30% fluorescently-labeled pre-tRNA. The labeled pre-tRNA was gel purified using 12% urea- PAGE and eluted using the crush-soak method (47). Purified products were concentrated using 10 kDa MWCO Amicon® Ultra Centrifugal Filters, and ethanol precipitation. Pre-tRNA stocks were resuspended in 10 mM Tris, pH 8.0 with 1 mM EDTA, quantified by absorbance, and stored at - 80 °C.
The extinction coefficient for the B. subtilis pre-tRNAAsp at 260 nM was experimentally determined to be 674,390 M-1 cm-1 through alkaline hydrolysis. Concentrations of fluorescein were measured at 492 nm (extinction coefficient = 78000 M-1 cm-1). Prior to all assays, substrates were
thawed, diluted with nuclease-free water, and heated at 95 °C for 90 sec followed by refolding by incubating at 25 °C for ≥ 15 min, and then incubating with metal-containing buffer for ≥ 15 min.
Fluorescence anisotropy binding assays
Binding assays were performed in Corning black polystyrene half-area 96-well plates (Product number 3686), as previously described (22,46). In short, PRORP1 variants were serially diluted from 20 μM to 9 nM and equal volumes of enzyme and 20 nM 5´-fluorescein-pre-tRNA substrate were mixed; a minimum of 12 concentrations was analyzed. Enzyme-substrate mixtures were incubated at 25 ± 1 °C in 30 mM MOPS, pH 7.8, 330 mM NaCl, 1 mM TCEP, and 20 mM CaCl2. Anisotropy readings of the 5´-fluorescein-pre-tRNA tag were measured with a Tecan Ultra plate reader using an excitation wavelength of 485 nm, and emission wavelength of 535 nm. The anisotropy measurements at each enzyme concentration were observed 5 times over 15-20 min to ensure complete equilibration. The concentration dependence of the anisotropy changes was well described by a single binding isotherm (Equation 1) (where FA is the fluorescence anisotropy, FA0 is the initial anisotropy, ΔFA is the total change in anisotropy, P is the concentration of PRORP and KD is the dissociation constant). The KD values and standard error for KD values were
calculated by fitting Eq. 1 to the data points from a single experimental trial using GraphPad Prism to carry out non-linear regression analysis.
Equation (1) 𝐹𝐴 = 𝐹𝐴0+[𝑃]+𝐾𝐷∆𝐹𝐴∙[𝑃]
Single-turnover kinetic assays
Single-turnover assay reactions were initiated through addition of 5–45 µM enzyme to an equal volume of 30 nM 5´-fluorescein-pre-tRNA substrate. Reactions were carried out at 25 ± 1 °C in 30 mM MOPS, pH 7.8, 330 mM NaCl, 1 mM TCEP, and 20 mM MgCl2. At specific time
points (0–4800 sec), 4 µL aliquots of the reaction were quenched with an equal volume of 100 mM EDTA, 8 M Urea, 0.05% bromophenol blue, 0.05% xylene cyanol, and 2 µg/µL yeast tRNA. Fluorescently labeled pre-tRNA substrate and 5´ leader product were separated by electrophoresis on 22.5% denaturing urea-PAGE gel. Gels were visualized using an Amersham Typhoon Biomolecular Imager, and the fraction of product was quantified using ImageJ software. A minimum of 10 time points was analyzed for each mutant. Observed single-turnover rate constants and standard errors were obtained by fitting a single exponential to the data points from a single experimental trial using GraphPad Prism 8 (Equation 2).
Equation (2) 𝐹𝑟𝑎𝑐𝑡𝑖𝑜𝑛 𝑝𝑟𝑜𝑑𝑢𝑐𝑡 = 𝐴 − 𝐵(𝑒−𝑘𝑡)
Electrophoretic mobility shift assays
tRNAs were radiolabeled at the 5´ end using [γ-32P] ATP and T4 polynucleotide kinase, then purified using an Illustra MicroSpin G-25 column (GE Healthcare). RNA-binding reactions included 0.9 nM radiolabeled RNA and protein serially diluted (2-fold) from 25 µM to 3 nM. Binding reactions were incubated for 1 h at 20 °C in 20 mM HEPES, pH 7.9, 100 mM NaCl, 1 mM TCEP, 10 mM CaCl2, 0.02% (v/v) Tween 20, 0.1 mg/mL poly r(U) , and 2.5% (v/v) glycerol and separated by electrophoresis on 10% TBE polyacrylamide gels (Invitrogen). Gels were dried and exposed to storage phosphor screens for 6–20 h, scanned with a Typhoon 8600 Imager, and the band intensities were quantified with ImageQuant 5.2. The data for three technical replicates were analyzed and KD values were calculated via non-linear regression analysis for one-site binding with GraphPad Prism 7.
3.4 Results
Engineering a PRORP1 PPR domain for crystallization
To understand substrate recognition by PRORP enzymes, we sought to determine a crystal structure of a PRORP in complex with tRNA. Through protein engineering, we obtained crystals suitable for structure determination of the PPR domain of A. thaliana PRORP1 in complex with tRNAPhe. Our attempts to crystallize a PRORP containing both PPR and catalytic domains in complex with tRNA were unsuccessful. We therefore focused on the PRORP1 PPR domain as the module that drives tRNA recognition (24,34-36). We engineered three regions of the PPR domain to promote crystallization. First, we noted that in the crystal structure of Arabidopsis PRORP1 (residues 76-572, PDB ID: 4G24) helices α10 and α11 of the PPR domain form a hydrophobic interface with the central zinc-finger domain (22). Three aromatic residues within the PPR domain (Tyr266, Phe284 and Phe291) are located at the interface. To increase the solubility of the PPR domain, we substituted these aromatic residues with hydrophilic residues (Y266N, F284Q and F291Q). Second, previous studies demonstrated that the N-terminal flexible region is involved in tRNA binding and lysine residues in this region might contact tRNA (34,40), although residues 76-94 were disordered in the crystal structure of PRORP1. We defined the minimal N-terminal region required for tRNA binding by measuring binding of N-terminal deletions of the PPR domain (named N76, N83 or N86 to indicate the N-terminal residue with all fragments extending to the C-terminal residue 294) to yeast tRNAPhe by electrophoretic mobility shift assay (EMSA). The affinity of the N86 variant for tRNA was considerably weaker than that of the N76 and N83 variants (Table 3-1), indicating that the N83 variant is optimal for tRNA binding affinity. Third, a long loop (LAEAATESSP) between the two α-helices of the PPR2 motif is shorter in the other Arabidopsis isoforms, PRORP2 and PRORP3. Substituting this loop with a shorter loop
(LASASS) had little effect on tRNA binding affinity (Table 1, PRORP1 PPR N83 ΔPPR2 loop). The engineered PRORP1 PPR domain (84-294) with three solubilizing substitutions (Y266N, F284Q and F291Q) and a shorter loop in the PPR2 motif produced well behaved protein that retained tRNA recognition and was used successfully for crystallization in complex with yeast tRNAPhe. For simplicity, we will refer to this as the PRORP1 PPR domain.
Table 3-1 Binding affinities of PRORP1 PPR domain variants
a Measured using electrophoretic mobility shift assays in 20 mM HEPES, pH 7.9, 100 mM NaCl, 1 mM TCEP, 10 mM CaCl2, 0.02% (v/v) Tween 20, 0.1 mg/mL poly r(U) , and 2.5% (v/v) glycerol at room temperature. The mean and standard error of the mean values for three technical replicates are reported.
Structure description
We determined a crystal structure of the Arabidopsis PRORP1 PPR domain in complex with yeast tRNAPhe at 2.85 Å resolution by molecular replacement (Table 3-2). Two independent crystallographic complexes were present in an asymmetric unit. The N-terminal 11 residues of the PPR domain (SRKAKKKAIQQ) are disordered in the crystal structure in complex with tRNAPhe, despite their importance for tRNA binding (Table 3-1). The basic residues of the N-terminal flexible region could interact non-specifically with negatively-charged phosphate groups of the tRNA backbone to enhance substrate binding affinity. The two complexes include chain A (PPR domain)-chain B (tRNAPhe) and chain C (PPR domain)-chain D (tRNAPhe) and are highly similar
PRORP1 PPR Domain Variant tRNA KD (nM)a
PRORP1 PPR N76 Yeast tRNAPhe 627 ± 55
PRORP1 PPR N83 Yeast tRNAPhe 281 ± 44
PRORP1 PPR N83 ΔPPR2 loop Yeast tRNAPhe 317 ± 92
PRORP1 PPR N86 Yeast tRNAPhe 3405 ± 338
overall (root mean square deviation [rmsd] value of 1.00 Å over 192 CA atoms in the protein, 1.36 Å over 1569 atoms in the tRNA, and 1.94 Å over 1352 main chain atoms in the complex). However, each PPR domain in the asymmetric unit binds its corresponding tRNAPhe in a slightly different manner. The chain D tRNA molecule appears to be influenced by crystal packing forces, resulting in the chain C-chain D complex lacking some interactions. Hence, we focus on the chain A-chain B complex to describe the PPR domain-tRNA interaction.
Table 3-2 Crystallographic Summary of PRORP1 PPR-tRNAPhe
The crystal structure of the PRORP1 PPR domain-tRNA complex revealed that the PPR domain undergoes conformational changes that place PPR motifs 1-4 in position to interact with the tRNA. The PRORP1 PPR domain comprises 5 consecutive PPR repeats and one additional C- terminal helix (Figure 3-4A). The PPR5 motif does not interact with the tRNA. Instead it may aid in positioning the PPR1-4 motifs for tRNA elbow recognition relative to the nuclease active site. As noted in the crystal structure of full-length PRORP1, the central linker domain interacts with
the PPR5 motif and the terminal α-helix of the PPR domain (22). PPR5 together with the central linker domain bridges the tRNA elbow recognition and catalytic domains. It also serves as a C- terminal cap to the PPR domain, which stabilizes the terminal α-helices in the PPR4 motif (48). When compared with the PPR domain in the structure of full-length apo PRORP1 (PDB ID: 4G24), the first three PPR repeats (PPR1, PPR2 and PPR3) are different in their configurational details. The tRNAPhe-bound PPR domain exhibits a more curved conformation (Figure 3-3B). With the PPR4 and PPR5 motifs aligned, the PPR3 motif in the complex is shifted away from the tRNA, whereas PPR1 is closer to the tRNA molecule. These changes allow PPRs 1-4 to interact with the tRNA, inducing a more extensive interaction surface than had been predicted. Conformational flexibility is a common feature of PPR proteins. Previous studies show that PPR proteins utilize considerable structural adaptability to bind to single-stranded RNA (32,49). In contrast, the overall structure of yeast tRNAPhe is unaltered by the binding of the PPR domain. The structure of yeast tRNAPhe is highly similar to the previously determined structure of the tRNA alone (PDB ID: 1EHZ, rmsd value of 1.56 Å over 1568 atoms).
Figure 3-2 The PRORP1 PPR domain changes conformation to recognize the tRNA ‘elbow’. (A) The tRNA elbow binds to PRORP1 near PPR motifs 1 and 2. An Fo-Fc omit electron density map contoured at 3.0σ is superimposed with the tRNA. (B) Superposition of the PPR domains of the tRNA-bound PRORP1 PPR domain (green) and tRNA-free PPR domain (pink, PDB ID: 4G24). PPR domains are shown with α helices as cylinders, and the tRNA is shown as a backbone trace. Shifts in conformation are highlighted by arrows.
Figure 3-3 The PRORP1 PPR domain recognizes the tRNA D and TψC loops.
(A) Ribbon diagram of the crystal structure of the PRORP1 PPR-tRNAPhe complex. PRORP1 PPR is shown in green with tRNA-interacting residues displayed as stick models. The tRNA is shown as a cartoon colored by region: acceptor stem (cyan), D stem loop (blue), anticodon stem loop (magenta), variable region (yellow), and TψC stem loop (orange). (B) Close-up view of the G19- C56 base pair accommodation pocket of the PRORP1 PPR domain. PRORP1 and tRNA are colored as in (A) with atom colors: oxygen (red), nitrogen (blue), phosphorus (orange), and sulfur (yellow). Dashed lines indicate interactions between PRORP1 and tRNA, and the G19-C56 and G18-ψ55 base pairs are indicated by transparent spheres.
The PRORP1 PPR domain nestles the tRNA elbow in a pocket formed by PPRs 1-4
The Arabidopsis PRORP1 PPR domain recognizes conserved features of the ‘elbow’ of the L-shaped tRNA, formed by the D and TψC loops (Figure 3-4). A ubiquitous G19-C56 tertiary interaction between the D and TψC loops is located at the tip of the tRNA elbow, and the PPR domain forms a pocket that accommodates the G19-C56 base pair (Figure 3-4B). Residues that contact the TψC loop are more conserved than those that contact the D loop (Figure 3-5). The G19
base is surrounded by residues from the PPR1 motif (Asp105, Met106, Ser108, and Lys109) and Tyr133 from the PPR2 motif (Figures 3-4B and 3-6). The OH-group of Tyr133 is hydrogen bonded to the N2 atom of the G19 base, potentially a base-specific interaction (Figure 3-6). The C56 base forms a stacking interaction with the phenol ring of Tyr140 in the PPR2 motif (Figures 3-4B and 3-6). Together these interactions appear to recognize the structure as well as the sequence of the conserved base pair. In addition to recognizing the G19-C56 base pair, PRORP1 interacts with the base pair between G18 (D loop) and the pseudouridine, ψ55 (TψC loop). The tRNAPhe in our structure was obtained from yeast, so it has 14 post-transcriptional modification sites (50). The guanidinium group of Arg210 in the PPR3 motif is hydrogen bonded to the O2 atom of the ψ55 base (Figures 3-4B and 3-6). The intercalation of G57 between these two base pairs forms the tRNA elbow’s structural core whose sequence and structure are probed by PRORP1.
Figure 3-4 PRORP1 uses a conserved surface to interact with the TψC loop of tRNA. The PRORP1 PPR domain is shown as a space-filling sphere representation with residues colored by degree of conservation. The ConSurf server was used to identify sequence homologs and project the degree of conservation on the structure of the PRORP1 PPR domain. The tRNA is shown as a cartoon colored by region: acceptor stem (cyan), D stem loop (blue), anticodon stem loop (magenta), variable region (yellow), and TψC stem loop (orange).
Figure 3-5 PRORP1 PPR-tRNA interactions.
Schematic representation of interactions between the PRORP1 PPR domain and yeast tRNAPhe. PRORP1 PPR residues are green, tRNA TψC-loop nucleotides are orange, and tRNA D-loop nucleotides are light blue. Circles represent tRNA phosphate groups (P). Dotted and double lines indicate hydrophilic and stacking interactions, respectively.
The PRORP1 PPR domain forms a variety of additional tRNA interactions using basic side chains (Figure 3-6). Two unpaired nucleotides in the D loop are recognized by lysine side chains. Lys101 in the PPR1 motif stacks with the G20 base (Figure 3-6). The ε-amino group of Lys109 in the PPR1 motif makes a hydrogen bond with the O4 atom of the 5,6-dihydrouridine (D) base, D17 (Figures 3-4B and 3-6). This appears to be a base-specific contact recognizing a modified nucleotide but it is also capable of recognizing the O4 atom of an unmodified uracil. The phosphate backbone of the TψC loop is contacted by two arginine residues. The guanidino groups of Arg184